Cross-Specialty Performance of Large Language Models: Insights from Dental Medicine and Allergy and Clinical Immunology Examinations

Résumé fourni par la source

This Zenodo deposit contains the supplementary data and materials underlying the analyses reported in the manuscript “Cross-Specialty Performance of Large Language Models: Insights from Dental Medicine and Allergy and Clinical Immunology Examinations.” The deposit comprises: Derived scoring data used for the statistical analyses comparing the performance of large language model (LLM)-based systems on self-assessment examination questions from the Swiss Federal Licensing Examination in Dental Medicine (SFLEDM) and the European Examination in Allergy and Clinical Immunology (EEAACI). The data support between-model and within-model comparisons for text-based and image-capable models across both specialties. Instruction templates used to prompt the evaluated LLM-based systems to answer A-type and Kprim-type self-assessment questions from the SFLEDM and EEAACI. The templates were applied consistently across the evaluated systems, as described in the Methods section of the manuscript. The deposited materials are intended to support transparency and reproducibility of the reported analyses. Data availability and restrictions: The original examination questions and corresponding answer keys are not included in this deposit because their redistribution is restricted by the terms of use of the Institute for Medical Education (IML), University of Bern. Accordingly, the deposited scoring data and supplementary materials should be interpreted in conjunction with the methods and descriptions provided in the associated manuscript.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.